Large-scale models do not adequately resolve the spatial variability that contributes to surface fluxes in the atmosphere and thus can underestimate surface fluxes considerably, especially in oceanic regions where data that can be assimilated are sparse or non-existent. One way of improving flux estimates in large scale models is to parameterize the contribution to the surface flux by the unresolved directional variability in the near surface wind field. In a series of sensitivity tests, we examine the impact of such a subgrid parameterization on the NCEP operational model at different resolutions. The parameterization tested is formulated as a resolution dependent velocity scale term that is adaptable to varying model resolution. It is based on synergistic use of space-borne and in-situ wind data sets. It accounts for subgrid variability of the wind direction to spatial scales of 3.5 km. Experiments at high and low model resolution are done and the results are examined for the effect on the large scale circulation for a number of diagnostic indicators.